What AI does here
Once a day, an AI system (Anthropic's Claude) looks for globally significant stories that at least two independent, reputable outlets on our source allowlist have already published.
It reads those published reports and writes an original article in Thai and a separate original article in English. The two versions are written for their own readers, not translated word for word. Every claim is attributed to the outlet that reported it, and the outlets are listed and linked under Sources.
Before publishing, every article must pass automatic checks: at least two allowlisted sources, a length limit, short and few direct quotes, no allegations against people unless a court, police or official statement made them, no medical, financial or investment advice, and, when source text is available, a check for copied wording.
No person reviews articles before they are published. Articles that pass the checks are published automatically.
We correct errors when they are reported. Corrections are listed on the Corrections page and on the corrected article.
What AI does not do
It does not report from the scene, interview anyone, or have access to information that is not already published by the outlets it cites.
It does not invent quotes, figures or images. Images are used only when they are our own YouTube thumbnails or openly licensed Wikimedia Commons files with credit; otherwise stories have no image.
AI can still make mistakes, such as misreading a source or mixing up a number. Always check the linked sources for anything important, and tell us when you find an error.